Skip to content

Gemini 4 Hydrogen ships as Google's fast agent model

Flat pricing across a full 1 million-token context and 2.5x speed gains over Gemini 3.8 Flash make Hydrogen the default fast tier for coding and agent workloads.

Source: Google

Google released Gemini 4 Hydrogen today, positioned as the fast, everyday tier beneath its frontier Carbon and Argon models and billed as the successor to Gemini 3.8 Flash. It’s live now in the Gemini API through AI Studio, the Gemini app, Google Search and Gemini for business, and for paying subscribers in Android Studio, Google Antigravity and Gemini in Sheets.

Pricing is flat across the model’s full 1 million token context, in and out: $1.50 per million input tokens, $7.50 per million output tokens. Google says Hydrogen scores 51 against Argon’s 50 on the Artificial Analysis Intelligence Index (54 at its highest reasoning setting), more than triples 3.8 Flash’s Terminal-Bench 4.1 score (57% versus 18.2%), runs 2.5 times faster than 3.8 Flash at default settings, and cuts fabricated claims by more than five times.

Why it matters: flat per-token pricing across a full 1M-token window removes the usual arithmetic of output costing far more than input, or rates resetting once you cross a context threshold. Paired with the speed and fabrication numbers, this reads as a straightforward swap-in for anyone running agentic or coding workloads on 3.8 Flash.

The caveat: free Gemini app users don’t get to pick Hydrogen directly, they only get it through automatic model routing when Google decides the prompt warrants it, and the benchmark figures are Google’s own.

Share

Related reading

More from The Wire

All briefs